Apple's AI-powered voicemail transcription service incorrectly transcribed a business call from a Land Rover garage to a 66-year-old woman, inserting inappropriate sexual references and profanity instead of the actual message about a car event invitation.
A 66-year-old woman from Dunfermline, Louise Littlejohn, received a voicemail on Wednesday from a Lookers Land Rover garage in Motherwell inviting her to an event. Apple's AI-powered voice-to-text service transcribed the message incorrectly, asking if she had been 'able to have sex' and calling her a 'piece of ****' instead of accurately converting the business call about a car event between the sixth and tenth of March. The BBC confirmed the original audio was a conventional business call from a garage employee inviting her to a new car event. The transcription error likely occurred due to telephone audio quality, background noise at the garage, and the caller reading from a script rather than speaking naturally. The reference to 'sex' appears to have been a mistranscription of 'sixth' when the caller mentioned the sixth of March. Professor Peter Bell from the University of Edinburgh noted that while Scottish accents may have contributed, the primary factors were audio quality and background noise. This incident follows other recent Apple AI transcription errors, including the word 'racist' being transcribed as 'Trump' and false AI-generated news headline summaries.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI that exposes users to harmful, abusive, unsafe or inappropriate content. May involve providing advice or encouraging action. Examples of toxic content include hate speech, violence, extremism, illegal acts, or child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed